The present invention relates to moving picture encoding and decoding techniques using motion compensating prediction, and particularly to techniques for encoding and decoding a motion vector used for motion compensating prediction.
Motion compensating prediction is used in typical moving picture compression encoding. The motion compensating prediction is a technique for dividing an image of interest into fine blocks, assuming a decoded image as a reference image, and generating, as a predictive signal, a reference image moved from the same positions as blocks of interest in the image of interest by the amount of motion in a motion direction indicated by a motion vector. The motion compensating prediction may be performed by one motion vector unidirectionally and may be performed by two motion vectors bidirectionally.
For the motion vector, a motion vector of encoded neighboring blocks to blocks to be processed is assumed as a motion vector predictor (which will be denoted as “vector predictor” simply), and a difference between the motion vector of the blocks to be processed and the vector predictor is found and the differential vector is transmitted as an encode vector, thereby enhancing a compression efficiency.
A block size of the motion compensating prediction is further reduced and is made more varied in MPEG-4AVC than in MPEG-2, thereby enhancing an efficiency in motion compensating prediction. On the other hand, the reduced block size causes an increase in motion vectors, and thus there is a problem on the amount of codes in an encode vector.
A motion vector of a left-adjacent block to blocks to be processed is simply assumed as a vector predictor in MPEG-2 (ISO/IEC 13818-2 Information technology—Generic coding of moving pictures and associated audio information: Video), but a median of the motion vectors of a plurality of neighboring blocks is assumed as a vector predictor in MPEG-4AVC, thereby enhancing an accuracy of the vector predictor and preventing an increase in codes of the encode vector (ISO/IEC 14496-10 Information technology—Coding of audio-visual objects—Part 10: Advanced Video Coding). Further, there is known a technique that enhances an encode efficiency in an encode vector by use of a motion vector of another encoded image in MEPG-4AVC.    [Non-patent document 1] ISO/IEC 13818-2 Information technology—Generic coding of moving pictures and associated audio information: Video    [Non-patent document 2] ISO/IEC 14496-10 Information technology—Coding of audio-visual objects—part 10: Advanced Video Coding
The methods described in both ISO/IEC 13818-2 Information technology—Generic coding of moving pictures and associated audio information: Video and ISO/IEC 14496-10 Information technology—Coding of audio-visual objects—Part 10: Advanced Video Coding have a problem that since only one vector predictor can be obtained, a prediction accuracy is low and an encode efficiency is not improved. The present inventors have tried to take a method using a plurality of vector predictor candidates, but in this case, indexes for identifying the vector predictor candidates need to be encoded, and thus they have recognized a problem on an increase in codes of the indexes.